One-Step Classifier Ensemble Model for Customer Churn Prediction With Imbalanced Class

Abstract

In customer churn prediction, an important yet challenging problem is the class imbalance of data distribution. After analyzing the disadvantages of the commonly used “two-step” methods, this study combines multiple classifiers ensemble technique, self-organizing data mining with cost-sensitive learning, and proposes one-step classifier ensemble model for imbalance data (OCEMI). For each test customer, it can adaptively select out the more appropriate one from the two kinds of dynamic ensemble approach: dynamic classifier selection (DCS) and dynamic ensemble selection (DES).Meanwhile, new cost-sensitive selection criteria for DCS and DES are constructed respectively to improve the classification ability for imbalanced data. The empirical results show that this strategy can be used to predict customer churn more effectively.